KR101700889B1 - Image coding apparatus, image coding method, image decoding apparatus, image decoding method, and computer readable storage medium - Google Patents

Abstract

A picture coding apparatus configured to divide an image into at least one slice each including a plurality of blocks and to encode each slice in units of blocks includes a first encoding unit configured to encode a block included in a first portion of the slice, And a second encoding unit configured to encode a block included in a second portion of the slice, and when the second encoding unit encodes the first block of the second portion, The first encoding unit refers to the first quantization parameter referred to by the first encoding unit when the first encoding unit encodes the first block of the first portion, which is the first quantization parameter provided to the slice as a value, And encodes the first block included in the second part.

Description

TECHNICAL FIELD The present invention relates to an image encoding apparatus, an image encoding method, an image decoding apparatus, an image decoding method, and a computer readable storage medium,

The present invention relates to a picture coding apparatus, a picture coding method, an image decoding apparatus, an image decoding method, and a program, and more particularly to a method for coding / decoding quantization parameters in an image.

H.264 / MPEG-4 AVC (hereinafter referred to as " H.264 ") is known as a coding method used for compression recording of moving pictures (ITU-T H.264 (03/2010) Advanced video coding for generic audiovisual services). According to H.264, the quantization parameter of each block can be made arbitrary by coding the difference of the quantization parameter from the block immediately before the current block with mb_qp_delta information.

This value is encoded using a conventional binary arithmetic coding scheme employed in H.264. Specifically, each syntax element such as the above-described mb_qp_delta information is binarized, and as a result, a binary signal is generated. For each syntax element, the occurrence probability is provided in advance as a table (hereinafter referred to as "occurrence probability table"). The above-described binary signal is arithmetic-coded based on the occurrence probability table. Each time the binary signal is encoded, the occurrence probability table is updated based on the statistical information indicating whether or not the encoded binary signal has the highest probability of occurrence.

In recent years, activities have been initiated to standardize higher-efficiency encoding technology as a successor to H.264, and Joint Collaborative Team on Video Coding (JCT-VC) has been established between ISO / IEC and ITU-T. JCT-VC standardizes a coding technique called High Efficiency Video Coding (hereinafter referred to as "HEVC").

In the standardization of the HEVC, various coding methods are extensively studied not only for improving the coding efficiency but also for other aspects such as ease of implementation and shortening of the processing time. In order to shorten the processing time, for example, methods for improving the parallelism on the assumption that a coding method is used in a multi-core CPU are being studied. One of these methods is the so-called "Wavefront" method for processing entropy encoding / decoding in parallel (JCT-VC magazine, Internet <http://phenix.int-evry.fr/jct/doc_end_user / documents / 6_Torino / wg11 /> available from JCTV-F274.doc). Since the next object to be encoded must be encoded using the updated occurrence probability table, the processing can not be performed in parallel unless the statistical information is reset. However, resetting the statistical information leads to a problem that coding efficiency deteriorates. On the other hand, Wavefront applies the occurrence probability table obtained at the time of completion of encoding to a plurality of predetermined number of blocks to the leftmost block of the next line, thereby coding the blocks in parallel on a line-by-line basis while suppressing the deterioration of the coding efficiency . This is mainly an explanation of the encoding process, but can also be applied to the decoding process.

However, although Wavefront can improve the parallelism of arithmetic coding / decoding of each line, in practice, quantization and dequantization can not be performed until the quantization parameter of the immediately preceding block is determined in raster scanning. Therefore, even in the current Wavefront embodiment, there is a problem that the entire encoding / decoding process can not be executed in parallel.

The present invention relates to enabling parallel encoding / decoding as a whole processing including quantization / dequantization processing when parallel encoding / decoding of blocks on a line basis using Wavefront method.

According to one aspect of the present invention, a picture coding apparatus configured to divide an image into one or more slices each including a plurality of blocks and to encode each slice in units of blocks includes a block included in the first part of the slice And a second encoding means configured to encode a block included in a second portion of the slice, wherein when the second encoding means encodes the first block of the second portion, Wherein the second encoding means includes a second quantization parameter storage means for storing a first quantization parameter which is provided to the slice as an initial value and in which when the first encoding means encodes the first block of the first portion, The first block included in the second part is encoded with reference to the quantization parameter.

According to an exemplary embodiment of the present invention, parallel encoding / decoding can be realized as a whole process including quantization / dequantization processing when parallel encoding / decoding of blocks is performed line by line using Wavefront method.

1 is a block diagram showing a configuration of a picture coding apparatus according to a first exemplary embodiment.
Fig. 2 shows the configuration of a block line.
3 is a flowchart showing a frame encoding process by the picture coding apparatus according to the first exemplary embodiment.
4 is a flowchart showing a process of coding an uppermost block line by the picture coding apparatus according to the first exemplary embodiment.
5 is a flowchart showing a coding process of a block line other than the uppermost block line by the picture coding apparatus according to the first exemplary embodiment.
6 is a flowchart showing a block encoding process by the picture coding apparatus according to the first exemplary embodiment.
Fig. 7A shows the transfer of quantization parameters by the conventional picture coding apparatus.
FIG. 7B shows transmission of quantization parameters by the picture coding apparatus according to the first exemplary embodiment.
8 is a block diagram showing a configuration of an image decoding apparatus according to the second exemplary embodiment.
9 is a flowchart showing frame decoding processing by the image decoding apparatus according to the second exemplary embodiment.
10 is a flowchart showing a decoding process of the uppermost block line by the picture decoding apparatus according to the second exemplary embodiment.
11 is a flowchart showing a decoding process of block lines other than the uppermost block line by the picture decoding apparatus according to the second exemplary embodiment.
12 is a flowchart showing a block decoding process by the image decoding apparatus according to the second exemplary embodiment.
13 is a block diagram showing an example of a hardware configuration of a computer usable as the picture coding apparatus and the picture decoding apparatus according to the exemplary embodiment of the present invention.
14 is a block diagram showing a configuration of a picture coding apparatus according to the third exemplary embodiment.
15 is a flowchart showing a process of coding an uppermost block line by the picture coding apparatus according to the third exemplifying embodiment.
16 is a flowchart showing a coding process of a block line other than the uppermost block line by the picture coding apparatus according to the third exemplary embodiment.
17A shows transmission of a quantization parameter by the picture coding apparatus according to the third exemplary embodiment.
17B shows the transfer of the quantization parameter by the picture coding apparatus according to the third exemplary embodiment.
18 is a block diagram showing a configuration of an image decoding apparatus according to the fourth exemplary embodiment.
19 is a flowchart showing a decoding process of the uppermost block line by the picture decoding apparatus according to the fourth exemplary embodiment.
20 is a flowchart showing a decoding process of a block line other than the uppermost block line by the image decoding apparatus according to the fourth exemplary embodiment.
21 is a block diagram showing the configuration of a picture coding apparatus according to the fifth exemplary embodiment.
22 is a flowchart showing a coding process of the upper block line according to the fifth exemplary embodiment.
23 is a flowchart showing a coding process of a block line other than the uppermost block line according to the fifth exemplary embodiment.
24 is a block diagram showing a configuration of an image decoding apparatus according to the sixth exemplary embodiment.
25 is a flowchart showing a decoding process of an uppermost block line according to the sixth exemplary embodiment.
26 is a flowchart showing a decoding process of a block line other than the uppermost block line according to the sixth exemplary embodiment.

Other features and aspects of the present invention will become apparent from the following detailed description of illustrative embodiments thereof, which proceeds with reference to the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features and aspects of the invention and, together with the description, serve to explain the principles of the invention.

Various exemplary embodiments, features and aspects of the present invention will now be described in detail with reference to the accompanying drawings.

1 is a block diagram showing a picture coding apparatus according to a first exemplary embodiment.

Referring to Fig. 1, the selector 101 determines whether or not a block to be processed belongs to an even block line. The selector 101 outputs the block to the first coding unit 102 if the block belongs to the even block line, and outputs the block to the second coding unit 103 otherwise.

The first and second encoding units 102 and 103 encode a block for dividing the input image into n × n pixels ("n" is a positive integer of 2 or more) in units of lines as shown in FIG. 2 ( These units correspond to the "first encoding means" and the "second encoding means" in the claims, and the "encoding block included in the first portion of the slice" and the "encoding block included in the second portion of the slice" . Hereinafter, the line of a block is referred to as a "block line ". The present embodiment will be described based on an example using two encoding units, but the present invention is not limited to this. Referring to FIG. 2, a section 201 represented by a rectangle drawn with a thin line represents a block, and a section 202 represented by a rectangle drawn by a thick line represents a block line. In addition, a block in the white region representing the even block line including the upper block line (0th block line) is encoded by the first encoding unit 102. [ A block in a diagonal area indicating an odd-numbered block line is encoded by the second encoding unit 103.

Each of the first and second encoding units 102 and 103 first generates a prediction error according to a prediction with reference to a pixel or another frame around the current block to be encoded and performs orthogonal transformation to generate a transform coefficient. Then, each of the first and second encoding units 102 and 103 determines a quantization parameter of the orthogonally transformed transform coefficient, and quantizes each transform coefficient to generate a quantization coefficient. Then, each of the first and second encoding units 102 and 103 binarizes each of the syntax elements including the quantization coefficient to generate a binary signal. Each occurrence element is provided with an occurrence probability in advance as a table (hereinafter referred to as "occurrence probability table"). The binary signal is arithmetic-coded based on the aforementioned occurrence probability table. Every time the binary signal is encoded, the occurrence probability table is updated using statistical information indicating whether or not the encoded binary signal has the highest probability of occurrence.

The first occurrence probability table storage unit 104 stores the occurrence probability table generated by the first encoding unit 102. [ Hereinafter, the occurrence probability table stored in the first occurrence probability table storage unit 104 is referred to as a "first occurrence probability table ".

The first quantization parameter storage unit 105 stores quantization parameters determined by the first encoding unit 102. [ Hereinafter, the quantization parameter stored in the first quantization parameter storage unit 105 is referred to as "first quantization parameter ".

The second occurrence probability table storage unit 106 stores the occurrence probability table generated by the second encoding unit 103. [ Hereinafter, the occurrence probability table stored in the second occurrence probability table storage unit 106 is referred to as a "second occurrence probability table ".

The second quantization parameter storage unit 107 stores the quantization parameters determined by the second encoding unit 103. [ Hereinafter, the quantization parameter stored in the second quantization parameter storage unit 107 is referred to as "second quantization parameter ".

The integrated encoding unit 108 integrates the encoded data generated by the first encoding unit 102 and the encoded data generated by the second encoding unit 103 and outputs the combined data as a bit stream.

The operation of the picture coding apparatus according to the present exemplary embodiment will be described in detail with reference to flowcharts shown in Figs. 3 to 6. Fig. In the present exemplary embodiment, moving image data is input frame by frame, divided into blocks, and processed in raster order. Although the present exemplary embodiment is configured to input moving image data on a frame basis, it may be configured to input still image data corresponding to one frame, or to input image data on a slice-by-slice basis . For the sake of simplicity of explanation, the present embodiment will be described based on only the intra predictive encoding process, but the present invention is not limited to this. The present exemplary embodiment can also be employed in an inter prediction coding process.

First, in step S301, the picture coding apparatus determines whether the block to be processed belongs to the upper block line. If the block belongs to the upper block line (YES in step S301), the process proceeds to step S302. If the block does not belong to the upper block line (NO in step S301), the process proceeds to step S303.

The process of step S302 is a process of encoding the uppermost block line, and the details will be described later. The process of step S303 is a process for coding a block line other than the uppermost block line, and this process will be described in detail later. The selector 101 also determines whether the block line to which the block to be processed belongs is an even block line or an odd block line. If the block line is an even block line, the processing target block is independently encoded by the first encoding unit 102. If the block line is not an even block line, the processing target block is independently encoded by the second encoding unit 103.

Then, in step S304, the integrated encoding unit 108 integrates the encoded data output from the first encoding unit 102 and the encoded data output from the second encoding unit 103, And outputs it.

Then, in step S305, the picture coding apparatus determines whether or not all the block lines of the frame to be processed are coded. If all the block lines have been encoded (YES in step S305), the encoding processing of one frame ends. If all of the block lines are not coded (NO in step S305), the process returns to step S301 to start encoding of the next block line.

The processing in step S302 (processing for coding the upper block line) will be described in detail with reference to the flowchart shown in Fig. Since the upper block line is an even block line, the block to be processed is input to the first encoding unit 102 by the selector 101 and is encoded by the first encoding unit 102. [

First, in step S401, a quantization parameter serving as a reference of block coding is initialized to make the initial value of the quantization parameter equal to the slice. Hereinafter, a quantization parameter serving as a reference for block coding is referred to as a "block-based quantization parameter ". The quantization parameter used for quantizing the block to be encoded is not encoded as a syntax element as it is, and the difference value with the block reference quantization parameter is encoded. In this exemplary embodiment, this difference value corresponds to the cu_qp_delta value in the HEVC method. However, the present invention is not limited to this, and for example, the difference value described above may correspond to the mb_qp_delta code in the H.264 method. Then, in step S402, the occurrence probability table is initialized in a predetermined manner. The initialized occurrence probability table is used for arithmetic coding the first binary signal of the left end block of the block line, and is updated as necessary in step S403 described later. Hereinafter, the occurrence probability table used for arithmetic coding the first binary signal of the left end block of the block line will be referred to as "block line reference occurrence probability table ".

Then, in step S403, the first encoding unit 102 encodes the pixel data on a block-by-block basis.

In the present exemplary embodiment, one block is composed of 64 x 64 pixels, but the present invention is not limited to this. The size of the block may be as small as 32 x 32 pixels or as large as 128 x 128 pixels. The block encoding process in step S403 will be described in detail with reference to the flowchart shown in Fig.

First, in step S601, the first encoding unit 102 performs intraprediction on the input image block using pixels around the block, and generates a prediction error.

Then, in step S602, the first encoding unit 102 performs an orthogonal transformation on the prediction error to generate a transform coefficient. The first encoding unit 102 quantizes the transform coefficients using a quantization parameter (hereinafter referred to as a "block quantization parameter") determined based on, for example, the characteristics of the image and the code amount to generate a quantization coefficient do.

Next, in step S603, the first encoding unit 102 calculates a difference value between the block reference quantization parameter and the block quantization parameter to generate a cu_qp_delta value.

Next, in step S604, the first encoding unit 102 updates the block-based quantization parameter by setting the block quantization parameter used for encoding the block to be processed to the block-based quantization parameter. The block-based quantization parameter will be used to generate the cu_qp_delta value of the next block.

Next, in step S605, the first encoding unit 102 generates a binary signal by binarizing each of the syntax elements including the cu_qp_delta value and the above-described quantization coefficient. The first encoding unit 102 uses various binarization methods, such as unary binarization and fixed length binarization, in a similar manner to the H.264 method, for each syntax element. Also, the first encoding unit 102 performs arithmetic coding on the binary signal based on the occurrence probability table.

Then, in step S606, the first encoding unit 102 updates the occurrence probability table based on whether or not the arithmetic-coded binary signal has the highest probability of occurrence.

Then, in step S607, the first encoding unit 102 determines whether or not all the syntax elements in the block are arithmetic-coded. If all the syntax elements have been arithmetically encoded (YES in step S607), the block encoding processing is terminated. If all the syntax elements have not been coded (NO in step S607), the processing returns to step S605.

Referring again to FIG. 4, in step S404, the first encoding unit 102 determines whether or not the condition for storing the block reference quantization parameter is satisfied. In this exemplary embodiment, the condition for storing the block reference quantization parameter is whether or not the block coded in step S403 is the leftmost block of the block line. If the above condition is satisfied (YES in step S404), the process proceeds to step S405. In step S405, the block-based quantization parameter is stored in the first quantization parameter storage unit 105 as the first quantization parameter. If the above condition is not satisfied (NO in step S404), the process proceeds to step S406. The first quantization parameter will be used as the block reference quantization parameter when the second encoding unit 103 codes the left end block of the next block line.

Then, in step S406, the first encoding unit 102 determines whether or not the condition for storing the occurrence probability table is satisfied. In this exemplary embodiment, the condition for storing the occurrence probability table is whether the block coded in step S403 is a predetermined block from the left end block of the block line. If the above condition is satisfied (YES in step S406), the process proceeds to step S407. In step S407, the occurrence probability table is stored in the first occurrence probability table storage unit 104 as the first occurrence probability table. If the above condition is not satisfied (NO in step S406), the process proceeds to step S408. The first occurrence probability table will be used as a block line reference occurrence probability table when the second encoding unit 103 codes the left end block of the next block line.

Then, in step S408, the first encoding unit 102 determines whether or not all the blocks of the block line to be processed have been coded. If all the blocks have been coded (YES in step S408), the encoding of the uppermost block line ends. If all the blocks are not coded (NO in step S408), the process returns to step S403. In step S403, the next block is coded in raster order.

The processing of step S303 (processing for coding block lines other than the uppermost block line) will be described in detail with reference to the flowchart shown in Fig. The selector 101 determines whether or not the block line is an even block line for each block line. If the block line is an even block line, the image of the block line to be processed is input to the first encoding unit 102 and is encoded by the first encoding unit 102. If the block line is an odd block line, the image of the block line to be processed is input to the second encoding unit 103 and is encoded by the second encoding unit 103. First, the flow when the second encoding unit 103 encodes odd-numbered block lines will be described.

First, in step S501, a first quantization parameter is input as a block-based quantization parameter from the first quantization parameter storage unit 105. [ Then, in step S502, the first occurrence probability table is inputted from the first occurrence probability table storage unit 104 as a block line reference occurrence probability table.

The processing in steps S503, S504, S506, and S508 is similar to the processing in steps S403, S404, S406, and S408, and a description thereof will be omitted.

In step S505, the block-based quantization parameter is stored in the second quantization parameter storage unit 107 as the second quantization parameter. The second quantization parameter will be used as the block reference quantization parameter of the left end block of the next block line.

In step S507, the occurrence probability table is stored in the second occurrence probability table storage unit 106 as the second occurrence probability table. The second occurrence probability table will be used as a block line reference occurrence probability table when the left end block of the next block line is arithmetic-coded.

Next, the flow when the first encoding unit 102 codes the even block lines will be described.

First, in step S501, a second quantization parameter is input as a block-based quantization parameter from the second quantization parameter storage unit 107. [ Then, in step S502, the second occurrence probability table is input as the block line reference occurrence probability table from the second occurrence probability table storage unit 106. [

The processing in steps S503 to S508 is similar to the processing in steps S403 to S408, and a description thereof will be omitted here.

The above-described configuration and operation enable parallel coding by making it possible to refer to the occurrence probability table and the reference quantization parameter during the processing of the left end block, even before completing the processing of the block line just before the encoded block line. Figs. 7A and 7B each show an embodiment in which a block-based quantization parameter is referred to. According to the prior art, as shown in Fig. 7A, the processing of the next block line can not be started until the processing of the preceding block line is completed. However, according to the present exemplary embodiment, by allowing a reference to a block spatially above the left-end block to be processed, as a result of allowing a reference pattern as shown in Fig. 7B, There is no need to wait.

Further, in the present exemplary embodiment, when encoding the left end block, the quantization parameter used in the left end block of the block line directly above is used as the block reference quantization parameter. However, the present invention is not limited to this, and may be embodied in any configuration capable of improving the parallelism of processing in units of block lines. For example, the initial value of the quantization parameter provided to the slice can be used as a block reference quantization parameter when coding the left end block of all the block lines. In another possible configuration, the block reference quantization parameter may be the same as the condition for storing the occurrence probability table given in step S406 and step S506. Specifically, the quantization parameter used when coding the predetermined block from the left end block of the block line can be used as the block reference quantization parameter of the left end block of the next block line. Further, the picture coding apparatus may be configured to switch the block referred to as the block reference quantization parameter based on the coding mode of the left end block.

Further, in the present exemplary embodiment, arithmetic coding is used for entropy encoding, but the present invention is not limited to this. As long as the statistical information is used to entropy-encode the left end block of the next block line during the encoding of the block line when entropy encoding is performed based on the statistical information such as the occurrence probability table, an arbitrary encoding can be employed .

The present exemplary embodiment has been described based on an example using two encoding units. However, it is clear that, for example, by adding the third encoding unit, the third occurrence probability table storing unit, and the third quantization parameter storing unit, parallel processing by more encoding units becomes possible.

8 is a block diagram showing an image decoding apparatus according to the second exemplary embodiment.

Referring to Fig. 8, the selector 801 determines whether or not the block to be processed belongs to an even block line. The selector 801 outputs the above-mentioned bit stream to the first decoding unit 802 when the block to be processed belongs to the even block line, and outputs the bit stream to the second decoding unit 803 if not .

The decoding units 802 and 803 decode the input bit stream in units of block lines as shown in Fig. The present embodiment will be described based on an example using two decoding units, but the present invention is not limited to this. Referring to FIG. 2, a block of a white color area representing an even block line including an upper block line (0th block line) is decoded by the first decoding unit 802. FIG. The blocks in the diagonal area indicating the odd block lines are decoded by the second decoding unit 803.

Each of the first and second decoding units 802 and 803 first selects an occurrence probability table for the binary signal of the bitstream to be decoded and arithmetically decodes the binary signal based on the occurrence probability table to generate a quantization coefficient . Then, each of the first and second decoding units 802 and 803 inversely quantizes the quantization coefficient based on the quantization parameter to generate a transform coefficient. Each of the first and second decoding units 802 and 803 performs inverse orthogonal transformation on the transform coefficient to generate a prediction error. Next, each of the first and second decoding units 802 and 803 performs prediction with reference to the pixels or other frames around the decoding target block, and generates image data of the decoding target block. The first occurrence probability table storage unit 804 stores the occurrence probability table generated by the first decoding unit 802. [ The first quantization parameter storage unit 805 stores the quantization parameter determined by the first decoding unit 802. [

The second occurrence probability table storage unit 806 stores the occurrence probability table generated by the second decoding unit 803. [ The second quantization parameter storage unit 807 stores the quantization parameter determined by the second decoding unit 803. The image data integrating unit 808 shapes the image data generated by the first decoding unit 802 and the image data generated by the second decoding unit 803, and outputs the formed image data.

The operation of the image decoding apparatus according to the present exemplary embodiment will be described in detail with reference to flowcharts shown in Figs. 9 to 12. Fig. In the present exemplary embodiment, the bitstream is input frame by frame. The bitstream is divided into coded data fragments and then decoded. Although the exemplary embodiment of the present invention is configured to input a bit stream in frame units, it may be configured such that a frame is divided into slices and a bit stream is input in units of slices. For the sake of simplicity, the present embodiment will be described based on only the intra prediction decoding process, but the present invention is not limited to this. The present exemplary embodiment can also be employed in inter prediction decoding processing.

First, in step S901, the picture decoding apparatus determines whether the block to be processed belongs to the upper block line. If the block to be processed belongs to the upper block line (YES in step S901), the process proceeds to step S902. If the block to be processed does not belong to the upper block line (NO in step S901), the process proceeds to step S903.

The process of step S902 is a process of decoding the uppermost block line, and the details will be described later. The process of step S903 is a process for decoding block lines other than the uppermost block line, and the details of this process will be described later. Further, the selector 801 determines whether the block line to which the block to be processed belongs is an even block line or an odd block line. If the block line is an even block line, the processing target block is decoded independently by the first decoding unit 802. [ If the block line is not an even block line, the processing target block is independently decoded by the second decoding unit 803. In the present exemplary embodiment, the selector 801 determines whether or not the block line is an even block line, based on the number of decoded blocks. However, the present invention is not limited thereto. For example, the input bitstream may include an identifier provided in advance at the boundary between the block lines, and the selector 801 may determine, based on the identifier, whether the block line is an even block line. Alternatively, information indicating the size of the bit stream of each block line and the start position of the next block line may be provided, and the selector 801 may judge whether or not the block line is an even block line based on this information can do.

Next, in step S904, the image data integrating unit 808 integrates the image data output from the first decoding unit 802 and the image data output from the second decoding unit 803, and outputs the decoded image Output.

Then, in step S905, the picture decoding apparatus determines whether or not all the block lines of the processing target frame have been decoded. If all the block lines have been decoded (YES in step S905), the decoding processing of one frame ends. If all the blocks have not been decoded (NO in step S905), the process returns to step S901 to decode the next block line.

The processing in step S902 (processing for decoding the upper block line) will be described in detail with reference to the flowchart shown in Fig. Since the upper block line is an even block line, the coded data of the block line to be processed is input to the first decoding unit 802 by the selector 801 and decoded by the first decoding unit 802. [

Referring to Fig. 10, first, in step S1001, a quantization parameter serving as a reference for block decoding is initialized to make the initial value of the quantization parameter equal to the slice. Hereinafter, the quantization parameter serving as a reference for block decoding is referred to as a "block reference quantization parameter" as in the picture coding apparatus according to the first exemplary embodiment. The block quantization parameter when the decoding target block is inversely quantized is a state in which the value is not directly coded, but the difference value with the block reference quantization parameter is encoded as a syntax element. Therefore, at the time of decoding, a block quantization parameter should be generated by adding the block reference quantization parameter and the above difference value, and the decoder must perform inverse quantization using the generated block quantization parameter. In this exemplary embodiment, this difference value corresponds to the cu_qp_delta value in the HEVC method. However, the present invention is not limited thereto. For example, the difference value may correspond to the mb_qp_delta value in the H.264 method. Then, in step S1002, the occurrence probability table is initialized in a predetermined manner. The initialized occurrence probability table is used for arithmetic decoding the first binary signal of the left end block of the block line, and is updated as required in step S1003 described later. Hereinafter, the occurrence probability table used for the arithmetic decoding of the first binary signal of the first block of the block line is referred to as a " block line reference occurrence probability table "as in the picture coding apparatus according to the first exemplary embodiment.

Then, in step S1003, the first decoding unit 802 decodes the bit stream on a block-by-block basis to generate image data.

In the present exemplary embodiment, one block is composed of 64 x 64 pixels, but the present invention is not limited to this. The size of the block may be as small as 32 x 32 pixels or as large as 128 x 128 pixels. The block decoding process in step S1003 will be described in detail with reference to the flowchart shown in Fig.

First, in step S1201, the first decoding unit 802 performs arithmetic decoding on the bit stream based on the aforementioned occurrence probability table to generate a binary signal. Also, the first decoding unit 802 decodes the binary signal according to various binarization methods, such as unary binarization and fixed length binarization, in a manner similar to the H.264 method for each syntax element, Generate a containing syntax element.

Then, in step S1202, the occurrence probability table is updated based on whether or not the arithmetic decoded binary signal has the highest probability of occurrence.

Then, in step S1203, the first decoding unit 802 determines whether or not all the syntax elements in the block have been arithmetically decoded. If all the syntax elements have been arithmetically decoded (YES in step S1203), the processing proceeds to step S1204. If all the syntax elements have not been arithmetically decoded (NO in step S1203), the processing returns to step S1201.

Then, in step S1204, the first decoding unit 802 generates the block quantization parameter by adding the block reference quantization parameter described above and the decoded cu_qp_delta value in step S1201.

Then, in step S1205, the first decoding unit 802 generates a transform coefficient by inverse-quantizing the quantization coefficient based on the block quantization parameter. Then, the first decoding unit 802 performs inverse orthogonal transformation on the transform coefficient to generate a prediction error.

Next, in step S1206, the first decoding unit 802 updates the block-based quantization parameter by setting the block quantization parameter used in inverse-quantizing the block to be processed to the block-based quantization parameter. The block-based quantization parameter will be used to generate the block quantization parameter of the next block.

Then, in step S1207, the first decoding unit 802 performs intraprediction from pixels around the block to be processed to generate a predicted image. Further, the first decoding unit 802 adds the prediction error and the predicted image to generate image data corresponding to one block.

Referring back to the flowchart shown in Fig. 10, in step S1004, the first decoding unit 802 determines whether or not the condition for storing the block reference quantization parameter is satisfied. In this exemplary embodiment, the condition for storing the block reference quantization parameter is whether or not the decoded block in step S1003 is the left end block of the block line. If the above condition is satisfied (YES in step S1004), the process proceeds to step S1005. In step S1005, the block-based quantization parameter is stored in the first quantization parameter storage unit 805 as the first quantization parameter. If the above condition is not satisfied (NO in step S1004), the process proceeds to step S1006. The first quantization parameter will be used as the block reference quantization parameter when the second decoding unit 803 decodes the left end block of the next block line.

Then, in step S1006, the first decoding unit 802 determines whether the condition for storing the occurrence probability table is satisfied or not. In this exemplary embodiment, the condition for storing the occurrence probability table is whether or not the decoded block in step S1003 is a predetermined block from the left end block of the block line. If the above condition is satisfied (YES in step S1006), the process proceeds to step S1007. In step S1007, the occurrence probability table is stored in the first occurrence probability table storage unit 804 as the first occurrence probability table. If the above condition is not satisfied (NO in step S1006), the process proceeds to step S1008. The first occurrence probability table will be used as a block line reference occurrence probability table when the second decoding unit 803 decodes the left end block of the next block line.

Then, in step S1008, the first decoding unit 802 determines whether or not all the blocks of the block line to be processed have been decoded. If all blocks have been decoded (YES in step S1008), decoding of the uppermost block line is ended. If all blocks have not been decoded (NO in step S1008), the process returns to step S1003, from which the first decoding unit 802 decodes the next block in raster order.

The processing in step S903 (processing for decoding block lines other than the uppermost block line) will be described in detail with reference to the flowchart shown in Fig. The selector 801 determines whether or not the block line is an even block line for each block line. If the block line is an even block line, the bit stream of the block to be processed is input to the first decoding unit 802 and decoded by the first decoding unit 802. [ If the block line is an odd block line, the bit stream of the processing target block is input to the second decoding unit 803 and decoded by the second decoding unit 803. First, the flow when the second decoding unit 803 decodes an odd-numbered block line will be described.

First, in step S1101, a first quantization parameter is input as a block-based quantization parameter from the first quantization parameter storage unit 805. [ Then, in step S1102, the first occurrence probability table is input as the block line reference occurrence probability table from the first occurrence probability table storage unit 804. [

The processing in steps S1103, S1104, S1106, and S1108 is similar to the processing in steps S1003, S1004, S1006, and S1008, and a description thereof will be omitted.

In step S1105, the block-based quantization parameter is stored in the second quantization parameter storage unit 807 as the second quantization parameter. The second quantization parameter will be used as the block reference quantization parameter of the left end block of the next block line.

In step S1107, the occurrence probability table is stored in the second occurrence probability table storage unit 806 as the second occurrence probability table. The second occurrence probability table will be used as a block line reference occurrence probability table when the first decoding unit 802 performs arithmetic decoding on the left end block of the next block line.

Next, the flow when the first decoding unit 802 decodes the even block lines will be described.

First, in step S1101, a second quantization parameter is input as a block-based quantization parameter from the second quantization parameter storage unit 807. [ Then, in step S1102, the second occurrence probability table is inputted as the block line reference occurrence probability table from the second occurrence probability table storage unit 806. [

The processing in steps S1103 to S1108 is similar to the processing in steps S1003 to S1008, and a description thereof will be omitted here.

The configuration and operation described above make it possible to refer to the block reference quantization parameter as well as the occurrence probability table which is statistical information during the processing of the left end block before completing the processing of the block line just before the block line which is currently being decoded, Enabling execution. Figs. 7A and 7B each show an embodiment in which a block-based quantization parameter is referred to. According to the prior art, as shown in Fig. 7A, the processing of the next block line can not be started until the processing of the preceding block line is completed. However, according to the present exemplary embodiment, by allowing a reference to a block spatially above the left-end block to be processed, as a result of allowing a reference pattern as shown in Fig. 7B, There is no need to wait.

In this embodiment, when decoding the left end block, the quantization parameter used in the left end block of the block line directly above is used as the block reference quantization parameter. However, the present invention is not limited to this, and may be embodied in any configuration capable of improving the parallelism of processing in units of block lines. For example, the initial value of the quantization parameter provided to the slice can be used as a block reference quantization parameter when decoding the left end block of all the block lines. As another possible configuration, the condition for storing the block reference quantization parameter may be the same as the condition for storing the occurrence probability table given in step S1006 and step S1106. Specifically, the quantization parameter when decoding the predetermined block from the left end block of the block line can be used as the block reference quantization parameter of the left end block of the next block line. Further, the picture decoding apparatus may be configured to switch the block referenced as the block reference quantization parameter based on the coding mode of the left end block.

Further, in the present exemplary embodiment, arithmetic decoding is used for entropy decoding, but the present invention is not limited thereto. As long as entropy decoding is performed on the basis of the statistical information such as the occurrence probability table and the statistical information is used to entropy-decode the left end block of the next block line during decoding of the block line, an arbitrary decoding can be adopted .

The present exemplary embodiment has been described based on an example using two decoding units. However, it is apparent that, for example, by adding the third decoding unit, the third occurrence probability table storing unit, and the third quantizing parameter storing unit, parallel processing by more decoding units becomes possible.

14 is a block diagram showing a picture coding apparatus according to the third exemplary embodiment.

Referring to Fig. 14, the selector 1401 determines whether the block to be processed belongs to an even block line. The selector 1401 outputs the block to the first encoding unit 1402 if the block belongs to the even block line, and outputs the block to the second encoding unit 1403 if not.

The first and second encoding units 1402 and 1403 encode a block for dividing the input image into n × n pixels ("n" is a positive integer of 2 or more) in units of lines as shown in FIG. The present embodiment will be described based on an example using two encoding units, but the present invention is not limited to this. Referring to FIG. 2, a section 201 represented by a rectangle drawn with a thin line represents a block, and a section 202 represented by a rectangle drawn by a thick line represents a block line. In addition, a block in the white region representing the even block line including the upper block line (0th block line) is encoded by the first encoding unit 1402. [ The blocks in the diagonal area indicating the odd block lines are encoded by the second encoding unit 1403.

Each of the first and second encoding units 1402 and 1403 first generates a prediction error according to a prediction with reference to a pixel or another frame around the current block to be encoded, and performs orthogonal transformation to generate a transform coefficient. Next, each of the first and second encoding units 1402 and 1403 determines a quantization parameter of the transform coefficient that is orthogonally transformed, and quantizes each transform coefficient to generate a quantization coefficient. Then, each of the first and second encoding units 1402 and 1403 generates a binary signal by binarizing each of the syntax elements including the quantization coefficient. Each occurrence element is provided with an occurrence probability in advance as a table (hereinafter referred to as "occurrence probability table"). The binary signal is arithmetic-coded based on the aforementioned occurrence probability table. Every time the binary signal is encoded, the occurrence probability table is updated using statistical information indicating whether or not the encoded binary signal has the highest probability of occurrence.

The initial quantization parameter storage unit 1404 stores the initial values of the quantization parameters.

The first occurrence probability table storage unit 1405 stores the occurrence probability table generated by the first encoding unit 1402. [ Hereinafter, the occurrence probability table stored in the first occurrence probability table storage unit 1405 is referred to as a "first occurrence probability table ".

The second occurrence probability table storage unit 1406 stores the occurrence probability table generated by the second encoding unit 1403. [ Hereinafter, the occurrence probability table stored in the second occurrence probability table storage unit 1406 is referred to as a "second occurrence probability table ".

The integrated encoding unit 1407 integrates the encoded data generated by the first encoding unit 1402 and the encoded data generated by the second encoding unit 1403, and outputs the combined data as a bit stream.

The operation of the picture coding apparatus according to the present exemplary embodiment will be described in detail with reference to the flowcharts shown in Figs. 3, 15 and 16. Fig. In the present exemplary embodiment, moving image data is input frame by frame, divided into blocks, and processed in raster order. Although the present exemplary embodiment is configured to input moving image data on a frame basis, it may be configured to input still image data corresponding to one frame, or to input image data on a slice-by-slice basis . For the sake of simplicity, the present embodiment will be described based on only the intra prediction coding process, but the present invention is not limited to this. The present exemplary embodiment can also be employed in inter-prediction coding processing.

First, the processes of steps S301, S304, and S305 shown in FIG. 3 are the same as those of the first exemplary embodiment, and a description thereof will be omitted.

The processing of step S302 (processing for coding the upper block line) will be described in detail with reference to the flowchart shown in Fig. The block to be processed is input to the first encoding unit 1402 by the selector 1401 and is encoded by the first encoding unit 1402 because the upper block line is an even block line.

First, in step S1501, a quantization parameter serving as a reference for block coding is initialized and stored in the initial quantization parameter storage unit 1404 so as to make the initial value of the quantization parameter equal to the slice. Hereinafter, the quantization parameter serving as a reference for block coding is referred to as "block-based quantization parameter" as in the first exemplary embodiment. The quantization parameter used for quantizing the block to be encoded is not encoded as a syntax element as it is, and the difference value with the block reference quantization parameter is encoded.

Then, in step S1502, the first encoding unit 1402 reads the quantization parameter initialized from the initial quantization parameter storage unit 1404 as a block reference quantization parameter for encoding the left end block of the block line. Next, the processes in steps S1503 to S1507 are similar to the processes in steps S402, S403, and S406 to S408 shown in Fig. 4, respectively, and a description thereof will be omitted here.

However, in step S1504, the first encoding unit 1402 encodes the pixel data on a block-by-block basis.

The processing in step S303 (processing for coding a block line other than the uppermost block line) will be described in detail with reference to the flowchart shown in Fig. The selector 1401 determines whether or not the block line is an even block line for each block line. If the block line is an even block line, the image of the block line to be processed is input to the first encoding unit 1402 and is encoded by the first encoding unit 1402. If the block line is an odd block line, an image of the block line to be processed is input to the second encoding unit 1403 and is encoded by the second encoding unit 1403. First, the flow when the second encoding unit 1403 encodes an odd-numbered block line will be described.

First, in step S1601, a block reference quantization parameter for encoding the left end block of the block line is input from the initial quantization parameter storage unit 1404. [ Next, in step S1602, the first occurrence probability table is input from the first occurrence probability table storage unit 1405 as a block line reference occurrence probability table.

In step S1603, the second encoding unit 1403 encodes the pixel data on a block-by-block basis. The processing in step S1604 is similar to the processing in step S1505 shown in Fig.

In step S1605, the occurrence probability table is stored in the second occurrence probability table storage unit 1406 as the second occurrence probability table. The second occurrence probability table will be used as a block line reference occurrence probability table when the first encoding unit 1402 performs arithmetic coding on the left end block of the next block line.

The processing of step S1606 is similar to the processing of step S1507 shown in Fig.

Next, the flow when the first encoding unit 1402 encodes the even block line will be described.

First, in step S1601, a block reference quantization parameter for encoding the left end block of the block line is input from the initial quantization parameter storage unit 1404. [ Then, in step S1602, the second occurrence probability table is input from the second occurrence probability table storage unit 1406 as a block line reference occurrence probability table.

The processing in steps S1603 to S1606 is similar to the processing in steps S1504 to S1507, and a description thereof will be omitted here.

The configuration and operation described above make it possible to refer to the block reference quantization parameter as well as the occurrence probability table which is statistical information during the processing of the left end block before completing the processing of the block line just before the block line which is currently coded, Enabling execution. 17A and 17B each show an embodiment in which a block reference quantization parameter is referred to. 17A and 17B, "Slice QP" represents the initial value of the quantization parameter provided to the slice. According to the conventional technique, as shown in Fig. 17A, the processing of the next block line can not be started until the processing of the preceding block line is completed. However, according to the present exemplary embodiment, since the initial value of the quantization parameter provided to the slice can be referred to as a block reference quantization parameter for coding the left end block of the block line, There is no need to wait for the completion of processing.

Further, in the present exemplary embodiment, arithmetic coding is used for entropy encoding, but the present invention is not limited to this. As long as the statistical information is used to entropy-encode the left end block of the next block line during the encoding of the block line when entropy encoding is performed based on the statistical information such as the occurrence probability table, an arbitrary encoding can be employed .

The present exemplary embodiment has been described based on an example using two encoding units. However, it is clear that, for example, by adding the third encoding unit and the third occurrence probability table storage unit, parallel processing by more encoding units becomes possible.

18 is a block diagram showing an image decoding apparatus according to the fourth exemplary embodiment.

Referring to Fig. 18, the selector 1801 determines whether the block to be processed belongs to an even block line. The selector 1801 outputs the above-mentioned bit stream to the first decoding unit 1802 when the block to be processed belongs to the even block line, and otherwise outputs the bit stream to the second decoding unit 1803 .

The decoding units 1802 and 1803 decode the input bit stream in units of block lines as shown in Fig. The present embodiment will be described based on an example using two decoding units, but the present invention is not limited to this. Referring to FIG. 2, a block in the white area representing the even block line including the upper block line (0th block line) is decoded by the first decoding unit 1802. [ The blocks in the diagonal area indicating the odd block lines are decoded by the second decoding unit 1803.

Each of the first and second decoding units 1802 and 1803 first selects an occurrence probability table for the binary signal of the bitstream to be decoded and arithmetically decodes the binary signal based on the occurrence probability table to generate a quantization coefficient . Then, each of the first and second decoding units 1802 and 1803 dequantizes the quantization coefficient based on the quantization parameter to generate a transform coefficient. Each of the first and second decoding units 1802 and 1803 performs inverse orthogonal transformation on the transform coefficient to generate a prediction error. Then, each of the first and second decoding units 1802 and 1803 performs motion compensation with reference to a pixel or another frame around the decoding target block, and generates image data of the decoding target block. The initial quantization parameter storage unit 1804 stores the initial values of the quantization parameters. The first occurrence probability table storage unit 1805 stores the occurrence probability table generated by the first decoding unit 1802. [

The second occurrence probability table storage unit 1806 stores the occurrence probability table generated by the second decoding unit 1803. The image data integrating unit 1807 shapes the image data generated by the first decoding unit 1802 and the image data generated by the second decoding unit 1803, and outputs the formed image data.

The operation of the image decoding apparatus according to the present exemplary embodiment will be described in detail with reference to flowcharts shown in Figs. 9, 19, and 20. Fig. In the present exemplary embodiment, the bitstream is input frame by frame. The bit stream is divided into coded data fragments corresponding to one block, and then decoded. Although the exemplary embodiment of the present invention is configured to input a bit stream in frame units, it may be configured such that a frame is divided into slices and a bit stream is input in units of slices. For the sake of simplicity, the present embodiment will be described based on only the intra prediction decoding process, but the present invention is not limited to this. The present exemplary embodiment can also be employed in inter prediction decoding processing.

First, the processes in steps S901, S904, and S905 shown in FIG. 9 are the same as those in the second exemplary embodiment, and a description thereof will be omitted.

The processing in step S902 (processing for decoding the upper block line) will be described in detail with reference to the flowchart shown in Fig. Since the upper block line is an even block line, the encoded data of the block line to be processed is input to the first decoding unit 1802 by the selector 1801 and decoded by the first decoding unit 1802. [

First, in step S1901, a quantization parameter serving as a reference for block decoding is initialized and stored in the initial quantization parameter storage unit 1804 so as to match the initial values of the quantization parameters with respect to the slice. Hereinafter, the quantization parameter as a reference for block decoding is referred to as a "block-based quantization parameter" as in the image decoding apparatus according to the second exemplary embodiment. The block quantization parameter when the decoding target block is inversely quantized is a state in which the value is not directly coded, but the difference value with the block reference quantization parameter is encoded as a syntax element. Therefore, at the time of decoding, a block quantization parameter should be generated by adding the block reference quantization parameter and the above difference value, and the decoder must perform inverse quantization using the generated block quantization parameter.

Then, in step S1902, the first decoding unit 1802 reads the value from the initial quantization parameter storage unit 1804 as a block reference quantization parameter for decoding the left end block of the block line. Then, the processes in steps S1903 to S1907 are similar to those in steps S1002, S1003, and S1006 to S1008, respectively, and a description thereof will be omitted here.

Next, the processing of step S903 (processing for decoding block lines other than the uppermost block line) will be described in detail with reference to the flowchart shown in Fig. The selector 1801 determines whether or not the block line is an even block line for each block line. If the block line is an even block line, the bit stream of the block to be processed is input to the first decoding unit 1802 and decoded by the first decoding unit 1802. [ If the block line is an odd-numbered block line, the bit stream of the processing target block is input to the second decoding unit 1803 and decoded by the second decoding unit 1803. First, the flow when the second decoding unit 1803 decodes an odd-numbered block line will be described.

First, in step S2001, a block reference quantization parameter for decoding the left end block of the block line is input from the initial quantization parameter storage unit 1804. Next, in step S2002, the first occurrence probability table is input from the first occurrence probability table storage unit 1805 as a block line reference occurrence probability table.

In step S2003, the second decoding unit 1403 decodes the pixel data on a block-by-block basis. The processing in step S2004 is similar to the processing in step S1905, and a description thereof will be omitted here.

In step S2005, the occurrence probability table is stored in the second occurrence probability table storage unit 1806 as the second occurrence probability table. The second occurrence probability table will be used as a block line reference occurrence probability table when the first decoding unit 1402 performs arithmetic decoding on the left end block of the next block line.

The processing in step S2006 is similar to the processing in step S1907, and therefore, a description thereof will be omitted. Next, the flow when the first decoding unit 1802 decodes the even block lines will be described.

First, in step S2001, a block reference quantization parameter for decoding the left end block of the block line is input from the initial quantization parameter storage unit 1804. Then, in step S2002, the second occurrence probability table is inputted from the second occurrence probability table storage unit 1806 as a block line reference occurrence probability table.

The processing in steps S2003 to S2006 is similar to the processing in steps S1904 to S1907, and a description thereof will be omitted here.

The configuration and operation described above make it possible to refer to the block reference quantization parameter as well as the occurrence probability table which is statistical information during the processing of the left end block before completing the processing of the block line just before the block line which is currently being decoded, Enabling execution.

Further, in the present exemplary embodiment, arithmetic decoding is used for entropy decoding, but the present invention is not limited thereto. As long as entropy decoding is performed on the basis of the statistical information such as the occurrence probability table and the statistical information is used to entropy-decode the left end block of the next block line during decoding of the block line, an arbitrary decoding can be adopted .

The present exemplary embodiment has been described based on an example using two decoding units. However, it is clear that, for example, by adding the third decoding unit and the third occurrence probability table storing unit, parallel processing by more decoding units becomes possible.

21 is a block diagram showing a picture coding apparatus according to the fifth exemplary embodiment.

Referring to FIG. 21, the selector 2101 determines whether or not the block to be processed belongs to an even block line. The selector 2101 outputs the block to the first encoding unit 2102 if the block belongs to the even block line, and outputs the block to the second encoding unit 2103 if not.

The first and second encoding units 2102 and 2103 encode a block for dividing an input image into n × n pixels ("n" is a positive integer of 2 or more) in units of lines as shown in FIG. The present embodiment will be described based on an example using two encoding units, but the present invention is not limited to this. Referring to FIG. 2, a section 201 represented by a rectangle drawn with a thin line represents a block, and a section 202 represented by a rectangle drawn by a thick line represents a block line. In addition, a block in the white region indicating the even block line including the upper block line (0th block line) is encoded by the first encoding unit 2102. [ A block in the diagonal area indicating an odd block line is encoded by the second encoding unit 2103.

Each of the first and second encoding units 2102 and 2103 first generates a prediction error according to a prediction with reference to a pixel or another frame around the current block to be encoded, and performs orthogonal transformation to generate a transform coefficient. Then, each of the first and second encoding units 2102 and 2103 determines a quantization parameter of the transform coefficient that is orthogonally transformed, and quantizes each transform coefficient to generate a quantization coefficient. Then, each of the first and second encoding units 2102 and 2103 generates a binary signal by binarizing each of the syntax elements including the quantization coefficient. Each occurrence element is provided with an occurrence probability in advance as a table (hereinafter referred to as "occurrence probability table"). The binary signal is arithmetic-coded based on the aforementioned occurrence probability table. Every time the binary signal is encoded, the occurrence probability table is updated using statistical information indicating whether or not the encoded binary signal has the highest probability of occurrence.

The initial quantization parameter storage unit 2104 stores the initial values of the quantization parameters. The initial occurrence probability table storage unit 2105 stores the initial value of the occurrence probability table. The integrated encoding unit 2106 integrates the encoded data generated by the first encoding unit 2102 and the encoded data generated by the second encoding unit 2103, and outputs the combined data as a bit stream.

The operation of the picture coding apparatus according to the present exemplary embodiment will be described in detail with reference to the flowcharts shown in Figs. 3, 22 and 23. Fig. In the present exemplary embodiment, moving image data is input frame by frame, divided into blocks, and processed in raster order. Although the present exemplary embodiment is configured to input moving image data on a frame basis, it may be configured to input still image data corresponding to one frame, or to input image data on a slice-by-slice basis . For the sake of simplicity, the present embodiment will be described based on only the intra prediction coding process, but the present invention is not limited to this. The present exemplary embodiment can also be employed in inter-prediction coding processing.

The processing of steps S301, S304, and S305 shown in FIG. 3 is the same as that of the first exemplary embodiment, and a description thereof will be omitted.

The processing in step S302 (processing for coding the upper block line) will be described in detail with reference to the flowchart shown in Fig. The block to be processed is input to the first coding unit 2102 by the selector 2101 and is coded by the first coding unit 2102 because the upper block line is an even block line.

First, in step S2201, a quantization parameter serving as a reference of block coding is initialized to be stored in the initial quantization parameter storage unit 2104 in order to make the initial value of the quantization parameter equal to the slice. Hereinafter, the quantization parameter serving as a reference for block coding is referred to as "block-based quantization parameter" as in the first exemplary embodiment. The quantization parameter used for quantizing the block to be encoded is not encoded as a syntax element as it is, and the difference value with the block reference quantization parameter is encoded.

Then, in step S2202, the first encoding unit 2102 reads the quantization parameter initialized from the initial quantization parameter storage unit 2104 as a block reference quantization parameter for encoding the left end block of the block line. Then, in step S2203, the occurrence probability table is initialized in a predetermined manner and stored in the initial occurrence probability table storage unit 2105. [ The occurrence probability table stored in the initial occurrence probability table storage unit 2105 is used for arithmetic coding the first binary signal of the left end block of the block line, and is updated as necessary in step S2205 described later. Hereinafter, the occurrence probability table used for arithmetic coding the first binary signal of the first block of the block line is referred to as a "block line reference occurrence probability table" as in the first exemplary embodiment.

Then, in step S2204, the first encoding unit 2102 reads the quantization parameter initialized from the initial occurrence probability table storage unit 2105 as a block line reference occurrence probability table.

Next, the processes in steps S2205 to S2206 are similar to the processes in steps S403 and S408 shown in Fig. 4, respectively, and a description thereof will be omitted here. However, in step S2205, the first encoding unit 2102 encodes the pixel data on a block-by-block basis. Next, the processing of step S303 (the processing of coding the block lines other than the uppermost block line) will be described in detail with reference to the flowchart shown in Fig. The selector 2101 determines whether or not the block line is an even block line for each block line. If the block line is an even block line, an image of the block line to be processed is input to the first encoding unit 2102 and is encoded by the first encoding unit 2102. If the block line is an odd block line, the image of the block line to be processed is input to the second encoding unit 2103 and is encoded by the second encoding unit 2103. First, the flow when the second encoding unit 2103 encodes odd-numbered block lines will be described.

First, in step S2301, a block reference quantization parameter for encoding the left end block of the block line is input from the initial quantization parameter storage unit 2104.

Then, in step S2302, the value is input from the initial occurrence probability table storage unit 2105 as a block line reference occurrence probability table.

Then, in step S2303, the second encoding unit 2103 encodes the pixel data on a block-by-block basis. The processing of step S2304 is similar to the processing of step S2206 shown in Fig. Next, the flow when the first encoding unit 2102 encodes the even block line will be described. First, in step S2301, a block reference quantization parameter for encoding the left end block of the block line is input from the initial quantization parameter storage unit 2104. Then, in step S2302, the value is input from the initial occurrence probability table storage unit 2105 as a block line reference occurrence probability table.

The processing in steps S2303 and S2304 is similar to the processing in steps S2205 and S2206, and a description thereof will be omitted here.

The above-described construction and operation may be performed by using a value initialized as a block reference quantization parameter as well as an occurrence probability table which is statistical information during the processing of the left end block before completing the processing of the block line just before the block line immediately coded currently, Enables parallel execution. Further, in the present exemplary embodiment, arithmetic coding is used for entropy encoding, but the present invention is not limited to this. An arbitrary encoding method can be adopted as long as the statistical information is initialized at the beginning of the encoding process, the encoding process is performed using the statistical information, and the statistical information is updated every time the encoding process is performed.

The present exemplary embodiment has been described based on an example using two encoding units. However, it is apparent that, for example, when a third encoding unit is added, parallel processing by more encoding units becomes possible.

24 is a block diagram showing an image decoding apparatus according to the sixth exemplary embodiment.

Referring to Fig. 24, the selector 2401 determines whether or not the block to be processed belongs to an even block line. The selector 2401 outputs the block to the first decoding unit 2402 if the block belongs to the even block line, and outputs the block to the second decoding unit 2403 if not.

The first and second decoding units 2402 and 2403 decode the input bit stream on a line-by-line basis as shown in Fig. Hereinafter, the line of a block is referred to as a "block line ". The present embodiment will be described based on an example using two decoding units, but the present invention is not limited to this. Referring to FIG. 2, a section 201 represented by a rectangle drawn with a thin line represents a block, and a section 202 represented by a rectangle drawn by a thick line represents a block line. In addition, the block in the white area indicating the even block line including the upper block line (0th block line) is decoded by the first decoding unit 2402. [ The blocks in the diagonal area indicating the odd block lines are decoded by the second decoding unit 2403.

Each of the first and second decoding units 2402 and 2403 first selects an occurrence probability table for the binary signal of the bitstream to be decoded and arithmetically decodes the binary signal based on the occurrence probability table to generate a quantization coefficient . Then, each of the first and second decoding units 2402 and 2403 dequantizes the quantization coefficient based on the quantization parameter to generate a transform coefficient. Each of the first and second decoding units 2402 and 2403 performs inverse orthogonal transformation on the transform coefficient to generate a prediction error. Then, each of the first and second decoding units 2402 and 2403 performs prediction with reference to a pixel or another frame around the decoding target block, and generates image data of the decoding target block.

The initial quantization parameter storage unit 2404 stores the initial values of the quantization parameters. The initial occurrence probability table storage unit 2405 stores the initial value of the occurrence probability table. The image data integrating unit 2406 shapes the image data generated by the first decoding unit 2402 and the image data generated by the second decoding unit 2403, and outputs the formed data.

The operation of the image decoding apparatus according to the present exemplary embodiment will be described in detail with reference to flowcharts shown in Figs. 9, 25 and 26. Fig. In the present exemplary embodiment, the bitstream is input frame by frame. The bit stream is divided into coded data fragments corresponding to one block, and then decoded. Although the exemplary embodiment of the present invention is configured to input a bit stream in frame units, it may be configured such that a frame is divided into slices and a bit stream is input in units of slices. For the sake of simplicity, the present embodiment will be described based on only the intra prediction decoding process, but the present invention is not limited to this. The present exemplary embodiment can also be employed in inter prediction decoding processing.

The processing of steps S901, S904, and S905 shown in FIG. 9 is the same as that of the second exemplary embodiment, and a description thereof will be omitted.

The processing in step S902 (processing for decoding the upper block line) will be described in detail with reference to the flowchart shown in Fig. Since the upper block line is an even block line, the encoded data of the block line to be processed is input to the first decoding unit 2402 by the selector 2401 and decoded by the first decoding unit 2402. [

First, in step S2501, a quantization parameter serving as a reference for block decoding is initialized to be stored in the initial quantization parameter storage unit 2404 in order to match the initial values of the quantization parameters with respect to the slice. Hereinafter, the quantization parameter serving as a reference for block decoding is referred to as a "block reference quantization parameter" in the same manner as the second embodiment. The block quantization parameter when the decoding target block is inversely quantized is a state in which the value is not directly coded, but the difference value with the block reference quantization parameter is encoded as a syntax element. Therefore, at the time of decoding, a block quantization parameter should be generated by adding the block reference quantization parameter and the above difference value, and the decoding apparatus must perform inverse quantization.

Next, in step S2502, the first decoding unit 2402 reads the value from the initial quantization parameter storage unit 2404 as a block reference quantization parameter for decoding the left end block of the block line. Next, in step S2503, the occurrence probability table is initialized by a predetermined method and stored in the initial occurrence probability table storage unit 2405. [ The occurrence probability table stored in the initial occurrence probability table storage unit 2405 is used for arithmetic decoding the first binary signal of the left end block of the block line, and is updated as necessary in step S2505 described later. Hereinafter, the occurrence probability table used for the arithmetic decoding of the first binary signal of the first block of the block line is referred to as a "block line reference occurrence probability table" as in the second exemplary embodiment.

Then, in step S2504, the first decoding unit 2402 reads the value from the initial occurrence probability table storage unit 2405 as a block line reference occurrence probability table.

Next, the processes in steps S2505 and S2506 are similar to the processes in steps S1003 and S1008 shown in Fig. 10, respectively, and a description thereof will be omitted here.

However, in step S2505, the first decoding unit 2402 decodes the pixel data on a block-by-block basis.

The processing of step S903 (processing for decoding block lines other than the uppermost block line) will be described in detail with reference to the flowchart shown in Fig. The selector 2401 determines whether or not the block line is an even block line for each block line. If the block line is an even block line, the bit stream of the processing target block is input to the first decoding unit 2402 and decoded by the first decoding unit 2402. [ If the block line is an odd block line, the bit stream of the block line to be processed is input to the second decoding unit 2403 and decoded by the second decoding unit 2403. First, the flow when the second decoding unit 2403 decodes an odd-numbered block line will be described.

First, in step S2601, a block reference quantization parameter for decoding the left end block of the block line is input from the initial quantization parameter storage unit 2404.

Then, in step S2602, the value is input from the initial occurrence probability table storage unit 2405 as a block line reference occurrence probability table.

Then, in step S2603, the second decoding unit 2403 decodes the pixel data on a block-by-block basis. The processing of step S2604 is similar to the processing of step S2506 shown in Fig. Next, the flow when the first decoding unit 2402 decodes the even block lines will be described. First, in step S2601, a block reference quantization parameter for decoding the left end block of the block line is input from the initial quantization parameter storage unit 2404. Then, in step S2602, the value is input from the initial occurrence probability table storage unit 2405 as a block line reference occurrence probability table.

The processing in steps S2603 and S2604 is similar to the processing in steps S2505 and S2506, and a description thereof will be omitted here.

The above-described configuration and operation are similar to the above-described configuration and operation in that, even before the completion of the processing of the block line just before the block line currently being decoded, a value initialized as the block reference quantization parameter, Enables parallel execution.

Further, in the present exemplary embodiment, arithmetic decoding is used for entropy decoding, but the present invention is not limited thereto. Any decoding method can be adopted as long as the statistical information is initialized at the beginning of the decoding process, the decoding process is performed using the statistical information, and the statistical information is updated each time the decoding process is performed.

The present exemplary embodiment has been described based on an example using two decoding units. However, it is apparent that, for example, when a third decoding unit is added, parallel processing by more decoding units becomes possible.

The above-described embodiments have been described on the assumption that each of the processing units shown in Figs. 1, 8, 14, 18, 21, and 24 is implemented as a hardware device. However, the processing performed by each of the processing units shown in Figs. 1, 8, 14, 18, 21, and 24 may be implemented by a computer program.

13 is a block diagram showing an example of the hardware configuration of a computer usable as the image processing apparatus according to each of the above-described exemplary embodiments.

The central processing unit (CPU) 1301 controls the entire computer using the computer program and data stored in the random access memory (RAM) 1302 and the read only memory (ROM) 1303, And performs various processes described as being performed by the image processing apparatus according to the exemplary embodiment. That is, the CPU 1301 functions as each of the processing units shown in Figs. 1, 8, 14, 18, 21, and 24.

The RAM 1302 has an area for temporarily storing, for example, computer program and data loaded from the external storage device 1306 and data acquired from the outside via the interface (I / F) 1307. [ The RAM 1302 has a work area used when the CPU 1301 executes various jobs. That is, for example, the RAM 1302 may be allocated as a frame memory, or may provide various other areas as needed.

The ROM 1303 stores, for example, setting data of this computer and a boot program. The operation unit 1304 includes a keyboard and a mouse. A user of this computer can input various instructions to the CPU 1301 by operating the operation unit 1304. [ The display unit 1305 displays the processing result performed by the CPU 1301. [ Further, the display unit 1305 includes a display device such as a liquid crystal display, for example.

The external storage device 1306 is a large capacity information storage device typified by a hard disk drive device. The external storage device 1306 is a computer that allows the CPU 1301 to implement the functions of the respective units shown in Figs. 1, 8, 14, 18, 21, Save the program. The external storage device 1306 may also store image data as an object to be processed.

The computer program and data stored in the external storage device 1306 are loaded into the RAM 1302 as required by control by the CPU 1301 and processed by the CPU 1301. [ The I / F 1307 can be connected to a network such as a local area network (LAN) or the Internet, and other devices such as a projection device and a display device. The computer can acquire and send various information through the I / F 1307. [ The bus 1308 connects each of the above units to each other.

As the operation according to the above-described configuration, the CPU 1301 plays a central role in controlling the operation described with reference to the above-described flowcharts.

The present invention may be implemented by supplying a storage medium storing a code of a computer program capable of realizing the above-described functions to a system, and reading and executing the code of the computer program. In this case, the code of the computer program read from the storage medium realizes the functions of the above-described exemplary embodiment, and the storage medium storing the code of the computer program falls within the scope of the present invention. Alternatively, an operating system (OS) or the like running on the computer may implement the above-described functions by performing some or all of the actual processing based on the instructions of the computer program code. Such a case is also within the scope of the present invention.

Alternatively, the present invention may be embodied in the following embodiments. That is, the code of the computer program read out from the storage medium can be written in a memory provided in a function expansion card inserted in the computer or a function expansion unit connected to the computer, and based on the instruction of the code of the computer program, Or a CPU or the like provided in the function expansion unit performs part or all of the actual processing, thereby realizing the above-described functions. Such a case is also within the scope of the present invention.

When the present invention is implemented by the above-mentioned storage medium, the storage medium stores the code of the computer program corresponding to the above-described flowchart.

While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. The following claims are to be construed as broadly as embracing all such modifications and equivalent structures and functions.

Priority is claimed on Japanese Patent Application No. 2011-243940, filed on November 7, 2011, the entirety of which is incorporated herein by reference.

Claims (18)

A picture coding apparatus comprising:
A first encoding means configured to encode a block of a first line of a slice including a plurality of blocks;
And second encoding means configured to encode a block of a second line of the slice,
Wherein the second encoding unit is configured to encode a first block that is first encoded in the encoding order of the second line, a first block provided as a slice and first encoded in the encoding order of the first line, Encoding the first block of the second line with reference to a first quantization parameter referred to by the first encoding means to encode the second block, and encoding the second block different from the first block of the second line , And encodes the second block of the second line with reference to a quantization parameter used to encode a block preceding the second block of the second line. The method according to claim 1,
Wherein the first quantization parameter is provided as the slice as an initial value and is referred to by the first encoding means for encoding the first block of the first line. delete The method according to claim 1,
Wherein the second encoding means refers to the quantization parameter used for encoding the first block of the second line to encode the second block of the second line, And coding the block. The method according to claim 1,
Wherein the first encoding means encodes a difference value between a quantization parameter used for encoding the first block of the first line and the first quantization parameter,
And the second encoding unit encodes a difference value between the quantization parameter used for encoding the first block of the second line and the first quantization parameter. The method according to claim 1,
Wherein the first encoding means encodes a difference value between a quantization parameter used to encode the second block of the first line and a quantization parameter used to encode the block preceding the second block of the first line and,
Wherein the second encoding unit is configured to generate a difference value between a quantization parameter used for encoding the second block of the second line and a quantization parameter used for encoding a block preceding the second block of the second line In the picture coding apparatus. The method according to claim 1,
The second encoding means refers to the statistical information based on the result of entropy encoding performed on the predetermined block of the first line of the slice by the first encoding means to convert the first block of the second line into Wherein the picture coding apparatus comprises: As an image decoding apparatus,
A first decoding unit configured to decode a block of a first line of a slice including a plurality of blocks;
And second decoding means configured to decode a block of a second line of the slice,
Wherein the second decoding unit decodes the first block decoded first in the decoding order of the second line when the first block is decoded first in the decoding order of the first line, Decodes the first block of the second line with reference to a first quantization parameter referred to by the first decoding means to decode the second block and decodes the second block different from the first block of the second line And decodes the second block of the second line by referring to a quantization parameter used for decoding a block preceding the second block of the second line. 9. The method of claim 8,
Wherein the first quantization parameter is provided to the slice as an initial value and is referred to by the first decoding means for decoding the first block of the first line. delete 9. The method of claim 8,
Wherein the second decoding means decodes the second block of the second line by referring to the quantization parameter used for decoding the first block of the second line, A picture decoding apparatus for decoding a block. 9. The method of claim 8,
The first decoding means decodes the difference value between the quantization parameter used for decoding the first block of the first line and the first quantization parameter,
And the second decoding unit decodes the difference value between the quantization parameter used for decoding the first block of the second line and the first quantization parameter. 9. The method of claim 8,
The first decoding means decodes the difference value between the quantization parameter used for decoding the second block of the first line and the quantization parameter used for decoding the block before the second block of the first line, and,
Wherein the second decoding unit is configured to obtain a difference value between the quantization parameter used to decode the second block of the second line and the quantization parameter used to decode the block before the second block of the second line To the image decoding apparatus. 9. The method of claim 8,
The second decoding means refers to the statistical information based on the result of the entropy decoding performed on the predetermined block of the first line of the slice by the first decoding means so that the first block of the second line And decodes the decoded image. As a picture coding method,
A first encoding step of encoding a block of a first line of a slice including a plurality of blocks;
And a second encoding step of encoding a block of a second line of the slice,
Wherein the first block is encoded in a coding order of the first line and the first block is encoded in a coding order of the first line when coding the first block to be coded first in the coding order of the second line, When encoding the first block of the second line with reference to the first quantization parameter referred to in the encoding step and encoding a second block different from the first block of the second line, The second block of the second line is encoded with reference to a quantization parameter used for encoding a block preceding the second block of the second block. As an image decoding method,
A first decoding step of decoding a block of a first line of a slice including a plurality of blocks;
And a second decoding step of decoding a block of a second line of the slice,
And a decoding unit for decoding the first block decoded first in the decoding order of the second line and decoding the first block decoded first in the decoding order of the first line, In a case where the first block of the second line is decoded with reference to the first quantization parameter referred to in the decoding step and the second block different from the first block of the second line is decoded, The second block of the second line is decoded with reference to a quantization parameter used for decoding a block preceding the second block of the second block. A computer-readable storage medium storing computer-executable instructions for causing a computer to perform a picture coding method when loaded and executed by a computer,
A first encoding step of encoding a block of a first line of a slice including a plurality of blocks;
And a second encoding step of encoding a block of a second line of the slice,
Wherein the first block is encoded in a coding order of the first line and the first block is encoded in a coding order of the first line when coding the first block to be coded first in the coding order of the second line, When encoding the first block of the second line with reference to the first quantization parameter referred to in the encoding step and encoding a second block different from the first block of the second line, The second block of the second line is encoded with reference to a quantization parameter used to code a block preceding the second block of the second block. A computer-readable storage medium storing computer-executable instructions for causing a computer to perform an image decoding method when loaded and executed by a computer,
A first decoding step of decoding a block of a first line of a slice including a plurality of blocks;
And a second decoding step of decoding a block of a second line of the slice,
And a decoding unit for decoding the first block decoded first in the decoding order of the second line and decoding the first block decoded first in the decoding order of the first line, In a case where the first block of the second line is decoded with reference to the first quantization parameter referred to in the decoding step and the second block different from the first block of the second line is decoded, Wherein the second block of the second line is decoded with reference to a quantization parameter used to decode a block preceding the second block of the second block.

Images